This paper presents the mitochondrial genome assembly and annotation of an unknown appendicularian species collected from Japan. The findings contribute valuable data for future eDNA studies.

The paper includes a thorough analysis of the mitochondrial genome including annotation and phylogenetics analyses.

However, the paper is missing important metrics regarding the PacBio Sequel II sequencing results. Please address the following points  to make the article scientifically sound:

1. Add the following metrics regarding the PacBio Sequel II sequencing results: HiFi Reads, HiFi Yield (bp), HiFi Reads Length (Mean bp), HiFi Read Quality (median), HiFi Number of Passes (mean), HiFi Reads Length Summary, HiFi Read Length Distribution.

2. Also include the stat for the assembly itself including: Contig type (Primary vs haplotigs; polished contigs; max contigs length; median contig length, N50 Contig Length, Sum of Contig length.

3. PacBio assemblers obtain circular contigs which include the mitochondria DNA (and bacteria DNA). If circular contigs were not utilized to directly obtain mtDNA, a brief explanation of the rationale for this decision would be valuable.

Response:

Thank you for the comments. We have added the PacBio sequencing metrics to the supplementary files.

We have also added to the method section the method for the fragmentation step, which enables mtDNA sequencing on the PacBio platform. The mitochondrial genome presented here was produced with the NOVOLoci software in targeted assembly mode, which yielded a linear concatenate sequence which we circularised by hand.

These changes have been reflected in the manuscript as follows:

“The sequenced reads were assembled with NOVOLoci ( https://github.com/ndierckx/NOVOLoci) in targeted assembly mode.”

“DNA was fragmented using Megaruptor 3® (Diagenode, USA) using the Megaruptor 3 Shearing Kit (Diagenode, USA #E07010003) at speed 32 and purified using SMRTbell cleanup beads (Pacific Biosciences, USA #102-158-300). Quality control of obtained DNA was performed using the Femto Pulse System (Agilent, USA) and the Genomic DNA 165 kb Kit (Agilent, USA #FP-1002-0275). The sequencing was performed on a PacBio ® Sequel II sequencer (Pacific Biosciences, USA) using the Sequel II sequencing kit 2.0 (Pacific Biosciences, USA #101-820-200). The DNA size profile and sequencing metrics are available on Zenodo (doi: 10.5281/zenodo.14934254).”

I suggest to better stress in the introduction the need to sequence other larvacean representatives, not only for eDNA studies but also for evolutionary studies aiming at understanding the origin of chordates and tunicates, and to unravel the hidden biodiversity of Larvacea, a quite neglected taxon

Response:

Thank you for this comment. We have added the following statement in the introduction:

“In addition to that, mitochondrial DNA sequence of larvaceans can be useful in evolutionary studies to the origin of tunicates, as well as chordates, and as a means to understand the hidden biodiversity of Larvacea, which is a quite neglected taxon.”

Was the DNA extracted from a single specimen preserved in 99% ethanol or at -80 °C?

Response:

Thank you for this comment. We preserved the specimen in 99% ethanol at -80°C. We have added the following statement to the manuscript:

“Samples were preserved in 99% ethanol at -80°C prior to DNA extraction.”

In the Methods section, the Authors reports that "Quality control of obtained DNA was performed using Agilent 4200 TapeStation (Agilent, USA). The sequencing was performed on a PacBio® Sequel II sequencer (Pacific Biosciences, USA)." Further details should be added. In particular, the Authors should report the ng and quality (i.e., absorbance values, TapeStation fragment profile) of the extracted DNA, the used PacBio library and sequencing kits, and the PacBio protocol used. Indeed, I suppose it was used the PacBio ultra-low input protocol, and, if so, how was the DNA fragmented?

Response:

Thank you for pointing this out. We have revised the quality control method, which used the Femto Pulse system after fragmentation. We also added the fragmentation protocol to the method section. As for the profile of the DNA fragments, we have attached the Femto Pulse run report to the supplementary files on Zenodo. We have adjusted the text as follows:

“DNA was fragmented using Megaruptor 3 ® (Diagenode, USA) using the Megaruptor 3 Shearing Kit (Diagenode, USA #E07010003) at speed 32 and purified using SMRTbell cleanup beads (Pacific Biosciences, USA #102-158-300). Quality control of obtained DNA was performed using the Femto Pulse System (Agilent, USA) and the Genomic DNA 165 kb Kit (Agilent, USA #FP-1002-0275). The sequencing was performed on a PacBio ® Sequel II sequencer (Pacific Biosciences, USA) using the Sequel II sequencing kit 2.0 (Pacific Biosciences, USA #101-820-200). The DNA size profile and sequencing metrics are available on Zenodo (doi: 10.5281/zenodo.14934254).”

As for the phylogenetic tree, it seems that it was reconstructed using all 13 protein-coding genes, including those absent in larvaceans. Moreover, the tree was reconstructed from nucleotide, not from amino acid sequences, as is common when long evolutionary distances are analyzed, in order to avoid substitution saturation phenomena. I suggest reconstructing the tree also from amino acid sequences, with/without the protein-genes absent in larvaceans, so considering only positions without gaps and then to compare the obtained trees. Bayesian trees should also be reconstructed to confirm the obtained phylogeny

We used nucleotides instead of amino acids because we are still uncertain on how to translate the TGA codon in the Mesochordaeus and Bathochordaeus mitogenomes (Pichon et al., 2020), and worried that it may affect the results. Following the reviewer’s suggestion, we computed trees from amino acid sequences using the same IQ-TREE command line as for nucleotide sequences (as it selects the best models automatically), with or without the proteins that are not found in the appendicularian mitogenomes. In both cases the results were consistent with the nucleotide tree, by placing the new mitogenome in the same clade as O. longicauda, placing the appendicularians as a sister clade to thaliaceans except for D. nationalis, and identifying Ciona as a clade distinct from the other phlebobranchians. Unfortunately, the amino acid-based trees did not recover the expected relation between the three chordate sub-phyla nor between tunicate classes. We added a copy of these computations to our supplemental material for the reviewers and readers convenience, nevertheless we do not provide guarantee that the trees can be useful for other purposes than accountability of our work for this revision. We also did not compute Bayesian trees for this revision due to the lack of expertise in our team.

A well formatted "full version" of the tree of Figure 3 should be provided in the Supplementary Material

We have added the full version of the tree in the supplementary material under Figure S4.

Although out of the scope of the paper, a brief comment on the recovered tunicate phylogeny should be reported

We added the following: “Considering the tunicates phylogeny, the tree recovered clades for the free-living Appendicularia, Thaliacea, and for the sessile Stolidobranchia, Aplousobranchia and Phlebobranchia, from which it detached the Ciona genus as a separate clade. A single thaliacean species, Doliolum nationalis, grouped with the sessile tunicates, however this is not fully supported by bootstrap values. This computed tree suggests that targeted sampling and sequencing of additional doliolids and Ciona species may be useful to further clarify the phylogeny of tunicates classes and orders.”

The Zenodo Supplementary material contains a folder with the results of IQ-TREE and RaxML analyses: could you please comment if there were differences between the trees obtained with the two methods?

The two trees have very similar topologies (identical when folded like in Figure 3), and we chose the IQ-TREE one because it provides bootstrap values that are easy to interpret.

For not-users of the IQ-TREE software, I suggest explaining the meaning of the used command-line options and reporting that this software performs ML tree reconstructions

We changed the method section’s text to “-T AUTO (run as many CPU threads as possible) --runs 3 (perform 3 runs to assess model convergence) --polytomy (allow unresolved branches) --ufboot 1000 (1000 boostraps with the ultrafast method) -m MFP (automatic selection of the best model for maximum likelihood inference using the ModelFinder Plus method”

The Authors should clarify the legend of Figure 3, including the number of genes analysed, the length of the alignment, the method and model used for the tree reconstruction.

We changed the figure legend to “Phylogenetic tree computed by maximum likelihood inference on a ~13 kbp codon alignment of 13 mitochondrial genome protein-coding genes collected from publicly available aquatic chordate genomes”. We did not indicate the model as we used one partition per gene and the software did not chose the same model for each partition.

"Possible explanations for this result are heteroplasmy or that the mitogenome consists of concatemers ": please explain better why you hypothesize the existence of concatemers.

"Furthermore, these two haplotypes are supported by homopolymer variants at multiple different loci": a table with all variants between the two haplotypes should be reported in the Supplementary Material, particularly if Figure 2 shows only a portion of these variants.

I suggest making available in NCBI also the second assembled haplotype

Response:

Thank you for this comment. We have reanalyzed the sequences and we could not verify if the observed variation were true variants or caused by sequencing errors in homopolymers and therefore decided to remove the hypothesis of a second haplotype. The one haplotype is the one that is published online. We apologize for this error, and we confirmed that the current available sequence on NCBI is the correct one. We have also modified the text to remove the parts referencing the haplotypes.

"As the Oikopleura genus is polyphylic in our phylogenetic analysis and that of others,..." : the Authors should cite the references where the Oikopleura genus was already found as polyphyletic

Response:

Thank you for this comment. We have added the following references to support this statement:

Galt CP, Grober MS, Sykes PF. Taxonomic Correlates of Bioluminescence Among Appendicularians (Urochordata: Larvacea). The Biological Bulletin. 1985 Feb;168(1):125–34.